For GPU owners & AI developers
Whether you are a GPU owner who doesn't want strangers SSH'd into your PC—or a developer tired of paying hourly VM rates and stopped-instance storage fees—legit2compute works one verified AI job and one API call at a time.
| Full-machine rental marketplace | legit2compute provider node | |
|---|---|---|
| Access model | Renter gets shell/container access to your box | Signed job dispatch only; nobody logs into your PC |
| Using your own PC | Interrupting a rental breaks SLA / hurts reliability score | Away mode yields cleanly when you game or work |
| Setup | Dedicated host configuration, pricing bids, port setup | One legit2node binary + Docker on Windows or Linux |
| Personal perk | None | Your own legit2clip jobs run free on your own GPU |
| Renting P2P GPU VMs | legit2compute API | |
|---|---|---|
| Billing | Hourly wall-clock time + stopped storage & egress | Pay strictly per token or finished media job, no idle burn |
| Host disconnects | Training or serving stops if a host drops offline | Automatic job retry and managed always-on model endpoints |
| Privacy control | Depends on each host operator | Prompts never stored; Private switch locks requests to our own servers |
| Time to first call | Provision VM, pull weights, expose endpoint | Change one base URL in your OpenAI SDK (demo key included) |
On a machine-rental marketplace like Vast.ai, a renter leases your entire box by the hour—meaning dedicated Linux setup, open access, and reliability penalties if your power blinks or you want to use your PC. legit2compute sends individual containerized AI jobs while you are away (Windows) or headless (Linux). Nobody logs into your machine, and away mode pauses cleanly when you sit down to use it.
Instead of renting a GPU virtual machine by the hour, babysitting stopped-instance storage fees, and managing your own model weights, legit2compute gives you an OpenAI-compatible LLM API and pay-per-result media jobs (images from $0.004, video from $0.07/s) on one key. You pay only for tokens and finished outputs, never idle hours.
Prompts and outputs are never stored. For sensitive agent or business workloads, flipping the Private switch on your developer dashboard keeps every request strictly on servers we own—never community provider nodes—and every response carries an X-L2C-Privacy header confirming the tier.
Yes. On Linux, legit2node runs headless around the clock without away mode. Cards with 16 GB or more of video memory can also join our model-hosting waitlist to be invited as paid demand arrives.
Every paid job your card finishes is validated and credited to your ledger. Hosting free user demos for part of your online time raises your share from 5% to 25% and gives you more turns at paid work. We do not promise an hourly rate.
More guides: Salad alternative · rent out your GPU for AI · earn with your GPU · OpenAI-compatible API.
Free to join, paused whenever you're using it, and a bigger share of every paid job when it hosts free demos.